comparing student ability using test scores

[charlie wong]

Hello

Suppose I got exam scores for students from different classes of different grades in a school. I want to predict how the students will perform in future exams when they move up to a higher grade. Note that students mix will be shuffled when moving up to a class of higher grade. Also, exams for different classes of the same grade differ in content so the scores also reflect difficulty of an exam. Students take multiple exams in each grade. Given the history of exam-related data (scores, classes, classmates) of every student, is there any method to adjust the scores to account for the difficulties of exams? How to do it with Stata?

Thanks!

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Without knowing what the data look like and what your model looks like, it is hard to help you..

If everyone takes the same tests and you use each test as an independent variable, then you can ignore differences in difficulty. In general, the issue is not mean (i.e., difficulty) but rather variance on the tests.

If you have different students taking different tests and you somehow want to equate the tests, you might consider standardizing the test values before creating a score.